Dive Brief
- More than four in five consumers — 82% — said that data security matters most in their AI-powered shopping experience, according to a survey of 2,000 U.S. consumers from Synchrony and Oxford Economics released Wednesday.
- More than three-quarters — 77% — said that they want clarity in how their data is being used.
- More than half of consumers said they would trust an AI shopping agent provided by a brand they already deal with, including a tech company, retailer or brand, their own personal bank or general AI platform. Consumers are much more likely to trust AI assistants when the stakes are low, however, than letting it make an expensive purchase.
Dive Insight
Consumer trust in a brand is hard won. As brands launch AI shopping assistants, they are increasingly leveraging their hard-earned trust to encourage adoption. However, such deployments often involve at least a few mistakes along the way.
If the launch of the AI assistant is bumpy, the base level of trust between company and customers may provide some cushion, but it might also cause the customer to reevaluate the brand.
“If there is a failure and it’s just a tech issue and doesn't really go to the brand's integrity in the consumer's thinking, that's one thing,” Brad Cleveland, CX expert and author, told CX Dive. “But if they ask, ‘Can I trust what they're doing here?’ That's something else, and they would then look at the company's response.”
Consumers also want control over the shopping process and the ability to review and approve purchases. Three-quarters of respondents said they would stop or pause using AI if it bought the wrong items without first checking. Three in 5 said the same even if it just recommended a poor-quality product, the survey found.
And, three-quarters of respondents said they would shop less or stop shopping with a given brand after an AI recommends a poor product.
Of course, brands should do everything they can to avoid such errors. Companies should closely monitor transactions and make sure a human is in the loop if automation flags an unusual or high-risk action, according to Synchrony. Brands should also be able to reverse mistakes and provide a clear remediation path.
Other steps companies can take include erecting guardrails to prevent unauthorized data use, being clear in what data is used and how, and explaining why products are being recommended, Cleveland said.
For CX leaders considering the rollout of AI shopping capabilities, Cleveland recommends they be transparent in why they’re rolling out these capabilities and convey that they've been launched ultimately with customers' best interests in mind.
“Be willing to make changes as you discover things that need attention. Have a good way to get feedback on it and use that feedback,” Cleveland said. “These things are never one and done, so know that roll out is just step one, and there will be tuning as you go along.”
